OTR kills the Bee

[tittof]

Bitlbee dies with

Fatal signal received: 11. That's probably a bug.

I think it happens when you already have someone's fingerprint and he already talked OTR to you and then starts a session with you again.

Someone else reported this to be happening to him aswell on the irc channel (zuran?)

This bug is also present in older OTR-Patch Versions of Bitlbee.

rgds

[zuran]

strace of bitlbee pid while generating a crash

[zuran]

Managed to crash bitlbee while fiddling with OTR. GDB gave the following info:

Program received signal SIGSEGV, Segmentation fault. 0x08062fe5 in op_still_secure (opdata=0xa04f440, context=0xa2b6060, is_reply=0) at otr.c:660 660 if(context->active_fingerprint->trust[0]) (gdb) Continuing.

Program terminated with signal SIGSEGV, Segmentation fault. The program no longer exists. (gdb) bt No stack. (gdb) print context No symbol "context" in current context.

[pesco]

Replying to zuran:

Managed to crash bitlbee while fiddling with OTR. GDB gave the following info:

Program received signal SIGSEGV, Segmentation fault. 0x08062fe5 in op_still_secure (opdata=0xa04f440, context=0xa2b6060, is_reply=0) at otr.c:660 660 if(context->active_fingerprint->trust[0])

this looks like a race condition where you managed to disconnect OTR again (=> active_fingerprint==NULL) before a message about going secure got to be printed. can you remember if that is consistent with the kind of fiddling you were performing? ;)

[pesco]

ah, actually this was likely just an oversight, a similar problem had been fixed in op_gone_secure but not applied to op_still_secure. should be fixed with revision 800 of ​http://code.khjk.org/bitlbee/